Re: [Dr. Geo] A question about DrGeo and MVC

[Oscar Nierstrasz]

OK, thanks. Then my understanding is correct. I first thought I needed to hook somehow into the `DrGeo` class or a subclass, but now I think it is better to go full MVC and break all dependencies from the model to the view. I'll experiment this week and see if I can start to make some real progress.

(It seems the hardest part of any migration project is to understand the existing architecture, and whether you should somehow hook into it, or rather circumvent it.)

Cheers,

Oscar
---
oscar.nierstrasz@feenk.com
https://feenk.com/about/

On 28 Apr 2024 at 18:11 +0200, Hilaire Fernandes <hfern@free.fr>, wrote:

At once DrGeo was developed like that. One model, several views, down to the item views.

It proved to be unneeded and it added complexity ans slownessss. I still have some dependencies between the model and view that are hard to garbage collect in some situation and it slows down the image when I deal with thousands of models. I need to clean up.

But any way, I am more in the tendency to not use anymore dependencies in that situation.

beDevelopment beWorkstation are not related to the view, more access to the ressources on the disk.

I think you can still do as you wish: several view per model as you will set the dependencies between the view to the model from an initialize method in the viewl. The model should not know about the, it just triggers the event to who is listening. See DrGMathItem>change

--  
GNU Dr. Geo
http://gnu.org/s/dr-geo/
http://gnu-drgeo.blogspot.com/